A predictive energy management strategy with pre-emptive load shedding for an islanded PV-battery microgrid

A predictive energy management strategy for an islanded photovoltaic powered microgrid with battery storage is presented in this paper. The strategy incorporates generation forecasts built from on-line weather service models combined with local photovoltaic system specifications. An energy balance prediction based on the battery SOC, the forecasted generation, and the forecasted load is evaluated and used to identify upcoming outages and to initiate automated load shedding to extend the run time of the system. The technique is algorithmic in nature and well suited for implementation on a networked control platform. Simulated and accelerated experimental results are presented to validate the proposed technique.